Skip to content
View nkrivoshey's full-sized avatar
🎯
Focusing
🎯
Focusing
1 follower · 0 following

Block or report nkrivoshey

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
nkrivoshey/README.md
Typing SVG

LinkedIn  Email  Telegram

🧠 About Me

I build analytics that companies actually make decisions from — not dashboards that sit unread.

At Metropolitan Premium Properties (Dubai's largest premium real estate brokerage) I own the full analytics stack as the sole analyst: designed the PostgreSQL DWH from scratch, unified 3 CRMs (Bitrix24 + PropCRM + 1C), delivered the company's first management-level Power BI platform, and shipped end-to-end marketing attribution.

Results that matter:

  • 10% revenue uplift via Marketing ROI dashboard and budget reallocation
  • 90% productivity boost for QC team through reporting automation
  • 5× faster reporting turnaround — eliminated 80% manual processing
  • 80 Dubai locations covered by ML-powered real estate intelligence platform

Before Dubai, at Sberbank (Russia's largest bank, 70M+ customers): built ML/NLP pipelines detecting 1.5M+ compliance violations and flagging 300K+ bankruptcy breaches.

📍 Dubai, UAE → Open to Europe · UAE · USA · Remote

💼 Production Work

Proprietary systems — code is internal, results are public.

Project Description Stack Impact
Leads Analytics Platform Company's first management analytics system: PostgreSQL DWH unifying Bitrix24 + PropCRM + 1C, multi-layer MV pipeline, Power BI lead funnel dashboard with stage-to-stage conversion & responsible rollup PostgreSQL · Power BI · DAX · Python First real-time lead visibility for C-level
CRM Data Architecture & DWH Cross-CRM analytics foundation: ingestion design, canonical data model (star schema + snapshots), data lineage manual, safe MV refresh patterns PostgreSQL · SQL · ETL Single source of truth for 3 CRMs
Direct Marketing Attribution End-to-end attribution: Maestro campaigns → Bitrix24 leads → 1C closed deals via BigQuery; unlocked first campaign-level ROI analysis BigQuery · Power BI · SQL First cross-CRM ROI analysis
Dubai RE Market Intelligence 80-location price platform with ML 3-month forecasts + Telegram bot; adopted by brokers for live client conversations Python · Power BI · Telegram API 80 locations, daily ML forecasts
Compliance Detection @ Sberbank ML/NLP pipelines across SMS + call campaigns + client communications; graph-based process mining for operational delays Python · PySpark · XGBoost · NLTK 1.5M+ violations · 300K+ cases

📂 Open Source & Portfolio

Project Description Stack
Dubai Real Estate Market Analysis Market analysis of 80 Dubai property areas: price dynamics, demand patterns, investment clusters Python · Pandas · Matplotlib ⭐ 7
LinkedIn AI Agent Production AI pipeline: news → Claude Sonnet → Telegram approval → LinkedIn publish. Live on Railway Python · Claude API · APScheduler · Railway 🚀 prod
Extended Kalman Filter (C++) Moment-based Kalman Filter with BFGS/DFP/L-BFGS matrix approximations — original research C++ · Numerical Methods academic
ML Practice Notebooks Structured ML tasks: regression, classification, NLP, PyTorch — from EDA to tuned models Python · scikit-learn · XGBoost · PyTorch —

🛠️ Tech Stack

Analytics & SQL

Advanced SQL DAX Power Query Data Modeling

BI & Visualization

Power BI Tableau Matplotlib Google Analytics Excel

Databases & DWH

PostgreSQL BigQuery MySQL Oracle GreenPlum

Python & ML / Stats / NLP

Python pandas scikit-learn PyTorch NLP R AB Testing

Engineering, Cloud & Tooling

Airflow MLflow PySpark Docker GCP AWS Git REST API LLM Dev

📊 GitHub Stats

 

🎓 Education & Research

  • M.Sc. Artificial Intelligence & Data Analysis — NIU BelGU (2025–present)
  • B.Sc. Digital Economy & Data Analytics — RANEPA, Moscow (2020–2025, GPA 4.5/5)
    Thesis: Russian equity mutual fund performance (K-means + CAPM/Carhart/Treynor–Mazuy).
    Findings requested by NAUFOR and the Central Bank of Russia.
  • Speaker — RANEPA Scientific Seminar on Investment Analytics (07/2025)
  • Finalist — Yandex "I am a Professional" Olympiad, Economics & Mathematics (2022)

Certifications: DataCamp (Supervised ML · Bayesian Analysis · SQL · Time Series) · Stepik (SQL Window Functions · Interactive SQL)

Pinned Loading

  1. linkedin-agent linkedin-agent Public

    Production AI pipeline: news collection → Claude Sonnet → Telegram approval → LinkedIn auto-publish | Python · Railway · APScheduler

    Python

  2. dubai-real-estate-market-analysis dubai-real-estate-market-analysis Public

    Market analysis of 80 Dubai property areas: price dynamics, demand patterns, ML price forecasts | Python · Power BI

    Jupyter Notebook 8 4

  3. extended-kalman-filter-cpp extended-kalman-filter-cpp Public

    Forked from ShirshovDIM/Extended_Moment-based_Kalman_Filter

    Extended & moment-based Kalman Filter with BFGS/DFP/L-BFGS matrix approximations | C++ · Numerical Methods

    C++

  4. pytorch-image-classification pytorch-image-classification Public

    End-to-end image classification with PyTorch: CNN, training loop, evaluation metrics | Python · Deep Learning

    Jupyter Notebook 1

  5. sberbank-nlp-challenge sberbank-nlp-challenge Public

    Sberbank internship challenge: multiclass text classification pipeline | Python · NLP · XGBoost

    Jupyter Notebook

  6. ml-practice-notebooks ml-practice-notebooks Public

    Structured ML practice: regression, classification, NLP, PyTorch — from raw data to tuned models | Python · scikit-learn · XGBoost · PyTorch

    Jupyter Notebook